Clinical studies have suggested that over 80% of cases of dementia in subjects over age 85 are due to Alzheimer's disease (AD), but clinicopathologic studies which assess the etiology of dementia in the oldest-old (subjects over 85) as well as the pathology of cognitively normal elderly are lacking. Over the past decade, we have performed prospective clinicopathological studies on over 86 elderly subjects who had been assessed with yearly neuropsychological testing. Our data suggest that a majority of cases of dementia in the oldest-old are not due to AD. In this renewal application, we propose to continue our clinicopathologic studies to describe the neuropathology of normal elderly and to determine the pathologic substrate of dementia in the oldest-old. We will utilize the expanded population in the clinical core, and estimate that we will be able to complete an additional 85 clinicopathologic studies by 1997. To test the hypothesis that dementia is associated with synapse loss and white matter pathology we will perform quantitative measures of synapse- associated proteins and galactolipids in standardized sections of cortex and white matter. We will then measure the association of these pathological variables with clinical measures such as memory test scores and VIQ. Because sensitivity and specificity of diagnosis of dementia subtype is poor in elderly subjects, we will analyze our autopsy proven cases to determine which clinical signs or pattern of scores on standard and experimental neuropsychological tests will improve diagnostic accuracy. We will test the hypothesis that cardiac ischemia is associated with ischemic dementias by performing 24 hour ambulatory EKGs in controls and subjects with different types of dementia.